1 / 62

ILC Detector R&D as seen by the Detector R&D Panel 2005-2007 and beyond …

ILC Detector R&D as seen by the Detector R&D Panel 2005-2007 and beyond …. (was for 3 years a Panel of the World-Wide Study Organising Committee) Jean-Claude Brient, Chris Damerell, Ray Frey, Dean Karlen, Wolfgang Lohmann, Hwanbae Park, Yasuhiro Sugimoto, Tohru Takeshita, Harry Weerts

Download Presentation

ILC Detector R&D as seen by the Detector R&D Panel 2005-2007 and beyond …

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. ILC Detector R&Das seen by the Detector R&D Panel 2005-2007and beyond … (was for 3 years a Panel of the World-Wide Study Organising Committee) Jean-Claude Brient, Chris Damerell, Ray Frey, Dean Karlen, Wolfgang Lohmann, Hwanbae Park, Yasuhiro Sugimoto, Tohru Takeshita, Harry Weerts Chris Damerell (RAL) ILC Detector R&D Chris Damerell

  2. What is ILC? ILC Detector R&D Chris Damerell

  3. SC linacs 31.5 MV/m for 11 km delivering 500 GeV collision energy (gradient is a major R&D topic – Lutz Lilje) • Undulator-based positron source (current baseline) (major R&D topic - Jim Clarke) • Electrons and positrons have just one damping ring each (issues of electron cloud – major R&D topic – Andy Wolski) • Single IR, 14 mrad crossing angle • 2 detectors operating in push-pull [all the benefits of two detectors, other than a luminosity advantage] • Machine must be upgradeable to 1 TeV • 4-volume Reference Design Report plus companion document was published October 2007 – but design will continue to evolve in light of ongoing R&D ILC Detector R&D Chris Damerell

  4. 0.2 s 337 ns Bunch structure at the ILC 1 ms 2820 bunch crossings • Detector options: • Single bunch timing • Time-slicing of train (eg at 50 ms intervals, 20 slices) • Integrate signals through train, with relaxed readout during the inter-train period • No ‘right answer’, despite statement of one collaboration that they will ‘time-stamp everything’ • There’s a power advantage in partial or complete time integration – fine sensor granularity can compensate for pileup from multiple bunch crossings • Lower power enables reduced material budget – desirable for physics • There’s been a successful history of exploiting granularity/time resolution tradeoffs in ACCMOR and SLD collaborations • Contrast LHC, where single bunch timing is mandatory ILC Detector R&D Chris Damerell

  5. The three detector concepts • LDC and GLD have merged into ILD for the LOI and EDR phase • LOIs if ‘validated’ by the IDAG will progress to ‘light’ or ‘demonstrator’ or ‘practice’ EDRs in 2010 (or 2012?) • Detectors to be built depend on R&D that cannot be completed before ~2012 ILC Detector R&D Chris Damerell

  6. Do we need R&D for ILC detectors? • “After all the R&D for LHC detectors (operating in a more hostile environment), this should be more than enough” • WRONG! • To satisfy the very challenging ILC physics goals, we need detectors that nobody knows how to build • What is easy, relative to LHC: • Instantaneous particle fluxes • Required radiation tolerance • What is difficult, relative to LHC: • The need for extraordinary jet energy resolution and vertexing performance • Special opportunities, relative to LHC: • Observe complex physics processes almost at the Feynman diagram level } 1/R2 to inferno at LHC collision point ILC Detector R&D Chris Damerell

  7. e+ e- t tbar At first sight, a confusing spray of particles … ILC Detector R&D Chris Damerell

  8. Mark Thompson The miracle of PFA (or equivalent jet energy resolution) reveals the flow of energy from the quarks of the primary interaction However, this is still not enough information for full physics analysis .. ILC Detector R&D Chris Damerell

  9. ILC vs LHC vertex detector parameters Which is better – a Sherman tank or a Ferrari? Each has its uses … ILC Detector R&D Chris Damerell

  10. A physics example – e+ e- b bbar Joanne Hewett, Sabine Riemann • Need highly polarised electron beams (longitudinal polarisation) • Need extremely clean b-tag to distinguish from other q-qbar processes • Need vertex charge to distinguish between b and bbar jets, otherwise see folded distns • These capabilities were pioneered at SLC/SLD, and are unique to the LC technology • Reward will be sensitivity to new physics via ‘oblique corrections’, where direct observation is beyond the reach of both ILC and LHC (example of large EDs, with 2TeV scale parameter) • Another important example – if LHC finds the Higgs, is it the SM Higgs, SUSY Higgs, or what? Precision measurements of branching ratios by ILC will be needed to decide ILC Detector R&D Chris Damerell

  11. Quark charge determination from ‘vertex charge’ e+e- -> b bbar from e+e- -> q qbar sample by flavour tag based on vertex topology (SLD procedure) For the b jets, which are quarks and which anti-quarks? 96% b-jets 4% bbar jets e-L (e+R) 96% bbar-jets 4% b jets In this event, total charge in decay chain for the backward jet resolves the forward-backward ambiguity – procedure pioneered in, and unique to, SLD (much cleaner than measurement of ‘jet charge’) ILC Detector R&D Chris Damerell

  12. Reviews of ILC Detector R&D PURPOSE: • Improved communication leading to enhanced R&D programmes • Get representatives of all R&D groups together for face-to-face discussions • Engage world-leading consultants from outside the ILC community, who would surely provide new insights – they did! • Ideally, the committee report would do little more than document mutually agreed changes from each review “If you don’t have buy-in, you can’t effect change.” • The reality proved a bit more complicated, but also more productive, due mainly to fresh contributions from those consultants ----------------------------------- SCHEDULE: • 3-day reviews were included in the 2007 regional workshops: • Beijing (Feb ’07) Tracking • DESY (LCWS June ’07) Calorimetry • Fermilab (Oct ’07) Vertexing ILC Detector R&D Chris Damerell

  13. Tracking Review Committee • Panel members: Chris Damerell (chair), Dean Karlen, Wolfgang Lohmann, Hwanbae Park, Harry Weerts • External consultants:Peter Braun-Munzinger, Ioanis Giomataris, Hideki Hamagaki, Hartmut Sadrozinski, Fabio Sauli, Helmuth Spieler, Mike Tyndel, Yoshinobu Unno • Regional representatives: Jim Brau, Junji Haba, Bing Zhou • RDB chair: Bill Willis • Local tracking experts: Chen Yuanbo, Ouyang Chun • Admin support: Naomi Nagahashi, Maura Barone, Maxine Hronek, Xu Tongzhou ILC Detector R&D Chris Damerell

  14. We reviewed the LCTPC, CLUCOU, SiLC and SiD tracking R&D collaborations • We were extremely impressed by the R&D programmes of all these groups, in some cases with very limited resources • However, we concluded that we are currently far from the goals, for all tracking options • Building a tracking system with excellent performance for qp >7 degrees will be challenging. Never achieved before and feasibility is not yet demonstrated • Forward tracking has generally performed badly. We all know the solution (drastic reduction in material budget) but can this be achieved in practice? • We became convinced of the need to construct large prototypes (~1 m diameter), and operate them under ILC-like beam conditions in a 3-5 T field, to establish what performance will be achievable at ILC, both for central and forward tracking • Not all the R&D collaborations felt that this would be necessary ILC Detector R&D Chris Damerell

  15. Lessons from LHC (ATLAS) ILC goal ILC Detector R&D Chris Damerell

  16. A new idea – Silicon Pixel Tracker • The most serious concern of the committee was the material budget, particularly how badly this might degrade the forward tracking: • For TPC tracker, can the endplate thickness really be reduced to ‘well below 0.3 X0 possibly 0.1 X0’? Our expert consultants were extremely doubtful • Franco Grancagnolo’s drift chamber could probably be made pretty thin, but would it provide robust track finding for high energy jets? Detailed simulations needed • For a silicon strip tracker, everyone now agrees that the ‘momenter’ concept is flawed. Will 5 single-sided layers (barrel or disks) suffice, or will there be serious pattern recognition problems, for example for high energy jets containing long-lived Bs, necessitating more layers and hence more material? • Discussions with our consultants led to a new suggestion – a silicon pixel tracker (SPT) which could deliver excellent pattern recognition for tracks in high energy jets, with very little material over the full range of polar angles • A preliminary study of this idea by Konstantin Stefanov looked promising – and we have been able to make a bit of progress since … ILC Detector R&D Chris Damerell

  17. A pixel tracker provides far more information per layer, is entirely free of ghost hits, and has a proven record for excellent pattern recognition compared to microstrips in high multiplicity jet-like events (ACCMOR Collaboration, mid-1980s) 200 GeV ‘jets’, Clean pattern recognition by two pixel planes 1 and 2 cm beyond the IP ILC Detector R&D Chris Damerell

  18. A tracker made with monolithic pixel sensors could provide the thinnest layers (~50 mm Si plus support structure) and the maximum information per layer, hence require the smallest number of layers • A major challenge could be to make such a detector with sufficiently low power to preserve gas cooling • The suggestion to achieve this is to dispense with single-bunch time stamping and even time slicing over most of the angular coverage, relying on the ECAL to label each track with its bunch number in the train ILC Detector R&D Chris Damerell

  19. SiD tracker layout (silicon microstrips) • 5 barrels and 4 endcaps, total area = 70 m2 • Everyone (?) now accepts need for standalone trk finding in this subsystem • With 50 μmsquare pixels – 28 Gpix system • Low mass support, gas cooling • If each sensor is 8 cm  8 cm (2.6 Mpix): 11,000 sensors is total • Note: forward disks will need time stamping, due to high 2-photon bgd ILC Detector R&D Chris Damerell

  20. one of 11,000 sensors 8x8 cm2 Cutout view without endcaps • SiC foam support ladders, linked mechanically to one another along their length • 5 closed cylinders (incl endcaps, not shown) will have excellent mechanical stability • ~0.6% X0 per layer, 3.0% X0 total, over full polar angle range, plus <1% X0 from VXD system (goal) • Scale is in line with trends in astronomical wide-field focal plane systems by 2020 ILC Detector R&D Chris Damerell

  21. ILC Detector R&D Chris Damerell

  22. SPT technologies • All options aim for ~15 mm precision with binary readout of 50 mm pixels • Similar area coverage to ATLAS SPT, but 5000 times more channels, 30 times less power, 20 times less material. Is this feasible? • CCD – Konstantin Stefanov • Reasonably confident in 100% min-I efficiency, though it hasn’t been demonstrated • Total in-detector power dissipation ~600 W is fine for gas cooling • LSST (3.2 Gpixels) being prototyped by e2V, will be a valuable 10% demonstrator ILC Detector R&D Chris Damerell

  23. Pinned Photodiode (PPD + 4T) - Konstantin Stefanov RG OD RSEL TG p+ pinning implant n photodiode shielding p+ substrate (p+) • PPD IP offered since ~5 years ago by numerous foundries for imaging • Pinning implant results in fully depleted n layer • Charge transfer gate TG decouples charge collection from sensing, permitting correlated double sampling and low noise (10 e- ENC quoted) • Large area PPD pixels being developed at RAL • Possible problems with inefficient transfer induced by small potential fluctuations in the photodiode area ILC Detector R&D Chris Damerell Konstantin Stefanov

  24. Photogate (PG + 4T) - Konstantin Stefanov RG OD RSEL Transfer Gate Collection gate(s) n buried channel shielding p+ substrate (p+) • Charge transfer allows correlated double sampling and low noise (10 e- possible) • LCFI is developing the underpinning technology for the ISIS • Charge transfer is fast due to funnel action (next slide) • Possible problems with inefficient transfer due to barely buried channel and inter-gate gaps (consequences of developing a combined CCD-CMOS process) • Hope of success with Jazz Semiconductor – currently merging with Tower ILC Detector R&D Chris Damerell Konstantin Stefanov

  25. PG pixel – possible layout 50 mm Full-area graded-potential photogate (PG) Transfer gate (TG) Time slicing or stamping requires a deep p-well of size to be determined, to shield the electronics Funnel, thanks to Grzegorz Deptuch V5<V4<V3<V2<V1<VTG VOD is held between V1 and VTG 1 2 3 4 Very small sense diode (SD) linking to 3T cct inside the TG ring collected charge confined in pixel by channel stop 5 TG (ring) 20 mm n-channel Depletion edge Central p-well (~5 mm diameter) housing 3-T cct Interface between epi and p+ Note: Charge collection directly to TG contributes to the signal. Unwanted charge collection directly to the tiny sense diode may be negligible, so a shielding deep p-implant may not be needed ILC Detector R&D Chris Damerell

  26. Calorimetry Review Committee • Panel members: Jean-Claude Brient, Chris Damerell, Wolfgang Lohmann (chair), Ray Frey • External consultants:Marcella Diemoz, Andrey Golutvin, Kazuhiko Hara, Robert Klanner, Peter Loch, Pierre Petroff, Jm Pilcher, Daniel Pitzl, Peter Schacht, Chris Tully • Regional representatives: Junji Haba, Michael Rijssenbeek, Jan Timmermans • RDB chair: Bill Willis • Admin support: Martina Mende, Naomi Nagahashi ILC Detector R&D Chris Damerell

  27. Ch Grah ILC Detector R&D Chris Damerell

  28. Overview of the review • Two main categories: • Very forward calorimetry(precision luminosity, hermeticity, beam diagnosics) • FCAL Collaboration (15 groups) • Doing a great job, but need additional resources, specially in USA • General calorimetry(precise jet energy measurement in multi-jet events, DE = 30%sqrt(E) • PFA approach: CALICE collab (41 gps), SiDCAL collab (17 gps, some in CALICE) • Compensating calorimetry: DREAM collab (8 gps), Fermilab gp • We were not able to exclude either option: much more work is required (and we might eventually need both to do the physics: PFA in barrel and compensating calorimetry forward) ILC Detector R&D Chris Damerell

  29. Tasks of the Forward Region IP • Precise measurement of the integrated luminosity (ΔL/L ~ 10-4) • Provide 2-photon veto LumiCal ~150mrad • Provide 2-photon veto • Serve the beamdiagnostics • using beamstrahlung pairs BeamCal ~40mrad GamCal 5mrad • Serve the beamdiagnostics • using beamstrahlung photons Beamstrahlung Ch Grah Challenges: High precision, high occupancy, high radiation dose, fast read-out! ILC Detector R&D Chris Damerell

  30. Main technical recommendations (FCAL) • Impressive report – physics requirements and technical implications were clearly presented • Design of LumiCal and BeamCal well advanced – GamCal (BS monitor) studies are at an early stage • BeamCal sensor development profits from close collaboration with groups developing rad hard sensors for hadron machines, notably sLHC • Need increased funding for travel, for their dedicated US collaborators (even before FY08 disaster), and for system-level engineering ILC Detector R&D Chris Damerell

  31. PFA approach to jet energy measurement • Goal is to separate depositions from charged and neutral hadrons in the ECAL/HCAL system. This is particularly challenging in the core of jets • Challenge (‘confusion term’) increases with jet energy and with reduced polar angle ILC Detector R&D Chris Damerell Mark Thomson

  32. Impressive results based entirely on simulations. Can such performance be achieved in a real system? • If possible, obtain data from charged and neutral hadrons in ‘physics prototype’ calorimeter system, and use them in conjunction with simulation of ILC jets to create more realistic hit patterns in the calorimetry system, hence determine how well PFA will handle real ILC events • There has been progress since our review (Jose Repond, Rajendran Raja) in establishing practical conditions for calibration with tagged neutrals (neutrons, KL, even anti-neutrons) using the MIPP2 facility in MCentre at Fermilab. DAQ problems of concern previously can be overcome • Don’t wait forever for Fermilab to pay for the modest MIPP upgrades to do this. The push needs to come from the ILC detector community, via our new directorate • This programme requires a significant effort, but this is better than discovering in 2025 that the PFA approach was a poor second choice • The vertex detector and tracking systems can and probably will be upgraded during ILC running, but not the coil or calorimetry – we do need to get these right when experiments choose their technologies ILC Detector R&D Chris Damerell

  33. Main recommendations (PFA systems) • While extremely promising, all studies to date (beyond the early experience with ALEPH and SLD) are based on simulations, hence subject to considerable uncertainty • These are only the averageshower radii. There is much greater uncertainty in the shape variability between individual showers, involving different inelastic scattering processes • Simulations alone cannot be trusted. Given the need to disentangle hits from charged and neutral showers, data are desirable on both, in large-scale ‘physics prototypes’ to: • Establish the performance truly achievable with such a calorimetry system • Establish which HCAL sensor technology (scintillator, RPCs, etc) will give the best performance ILC Detector R&D Chris Damerell

  34. Compensating calorimetry option ILC Detector R&D Chris Damerell

  35. Promising test beam results • Make no attempt to resolve the particles in jet cores, within the calorimeter • Crystal EM section, with dual readout of scintillation and Cerenkov light by timing , followed by a hadronic section with dual readout by quartz and scintillator fibres • No longitudinalsegmentation, but SiPMs and local readout chips will permit excellent hermeticity. HCAL thickness can be 10l or more • Simulations indicate they could achieve DE = 20-25%sqrt(E) for isolated jets. Not clear yet how well their pfa (John Hauptman) will sortout the crosstalk in multi-jet events ILC Detector R&D Chris Damerell

  36. Main recommendations(compensating calorimetry) • PFA performance is expected to degrade in the forward region, where for t-tbar and much BSM physics, one or more jets will generally be directed • Cannot afford to let the tracking ‘go to hell in the forward region’ as in the past • Less spreading of charged tracks may also favour a hardware compensating calorimeter and and pfa approach • Before moving to a large scale prototype, the review recommended they investigate a number of concerns, some by simulations, others by lab tests • Their collaboration needs more people, and we encourage others to join. Their approach could prove to be the outright winner – we simply don’t know yet ILC Detector R&D Chris Damerell

  37. Vertexing Review Committee • Panel members: Chris Damerell, Hwanbae Park (chair) • External consultants:Yasuo Arai, Dave Christian, Masashi Hazumi, Gerhard Lutz, Pavel Rehak, Petra Riedler, Steve Watts • Regional representatives: Tim Bolton, Chris Damerell, (Junji Haba) • RDB chair: Bill Willis • Local vertexing experts: Simon Kwan, Lenny Spiegel • Admin support: Naomi Nagahashi ILC Detector R&D Chris Damerell

  38. ILC vertex detector – two main layout options ILC Detector R&D Chris Damerell

  39. Optimal geometry will depend on ladder-end details that are not yet defined for any technology ILC Detector R&D Chris Damerell

  40. VXD technologies • All NINE approaches aim for ~3 mm precision and <40 mm 2-hit resolution • Target material budget is ~0.1% X0 per layer • They vary from single-bunch time stamping to time integrating with special compensating features • List them in approximate order of adventurousness – one or two are more likely to be candidates for second generation upgrades ILC Detector R&D Chris Damerell

  41. FPCCD – Yasuhiro Sugimoto • CCD with 5 mm pixels, read out once per train; 20 times finer pixel granularity instead of 20 time slices • Pair bgd rejected by mini-vectors indicating track direction • Bgd rejection depends on closely spaced pairs of sensors through the barrel • All signal processing is column parallel at ends of ladder, beyond active area • Possible showstopper**: • real bgd rejection factor proves to be less than ~20 as simulated ** one example showstopper per project, all agreed by the project leaders ILC Detector R&D Chris Damerell

  42. CPCCD – Andrei Nomerotski • Fast readout of CCD aiming for 50 ms frame rate • Main novel features are column parallel readout, with bump-bond connections on 20 mm pitch to readout chip including amp, analogue CDS, ADCs, sparsification and memory • In addition, generating the high drive current necessitated the development of special driver chips • Possible showstoppers: • Unacceptable bulk of service electronics at ladder ends • Biggest threat is that full-scale ladders won’t be made, due to lack of support from the UK funding agency (STFC) ILC Detector R&D Chris Damerell

  43. CMOS MAPS (MIMOSA) – Marc Winter • 3T architecture, limited to NMOS transistors in pixel • Rolling shutter – ‘row parallel’ to get the required readout rate • Goal is 25 ms (40 frames) on inner layer. Larger pixels on outer layers. Former may be too conservative, latter may be too optimistic. Detailed simulations needed • Plan to use 10-20 sensors per ladder, due to yield considerations • Possible showstopper: • Frame-rate CDS, not robust against baseline drift and low fcy pickup ILC Detector R&D Chris Damerell

  44. DEEP n-well – Valerio Re • Full CMOS in pixel, collecting signal charge o nthe deep n-well that houses the NMOS transistors (triple-well process) • In-pixel data sparsification and time-stamping with 30 ms precision • Goal is ~15 mm pixels, so binary readout OK • CDS achieved by in-pixel time-invariant signal processing • Possible showstopper: • Fall short of full min-I efficiency due to charge collection to competing in-pixel n-wells ILC Detector R&D Chris Damerell

  45. CAP – Gary Varner • CMOS MAPS, with signal storage (after charge-to-voltage conversion) on in-pixel capacitors • Aim for time slice < 50 ms with >10 storage cells, but difficult to achieve performance with adequate noise performance • Needs fast shaping time to accept signal from last BX before the sample. Signals are referenced to a baseline established at start of train, so there is exposure to baseline drift • Possible showstopper: • Insufficient pickup immunity due to charge-to-voltage conversion during the noisy bunch train ILC Detector R&D Chris Damerell

  46. DEPFET – Laci Andricek • Signal charge stored on ‘internal gate’ – unique in-house technology • Complex design – as well as sensors, need steering chips along edge of ladder, and readout chips bump-bonded at ladder ends • Possible showstopper: • Failure to reach required readout rate with full system ILC Detector R&D Chris Damerell

  47. Chronopixels – Dave Strom • Goal is to time-stamp (single bunch) by pixel functionality that can fit into a 10 mm pixel (full CMOS wirh 45 nm design rules) • Deep p-well to shield the signal charge from the PMOS transistors • Binary readout will give sufficient precision • Possible showstopper: • Unacceptably high power dissipation ILC Detector R&D Chris Damerell

  48. Vertically integrated pixel detectors (SOI & 3D) – Ray Yarema • An impressive strategy to be liberated from the constraints of CMOS by developing tiered systems • Potential for data-driven systems with single-bunch time stamping, the ‘physicists dream’ • Plan is for very small pixels with binary readout, like the chronopixels • Problems from back-gate effect with first manufacturers (Lincoln Labs) but a potentially clean solution with Tezzaron (wafer fab by Chartered Semiconductos in Singapore) • Cu-Cu thermocompression bonding (also being developed by IBM, MIT, …) • Chartered currently process 1000 wafers/month • Possible showstopper: • 4 Gpixels may exceed the power limits for gas cooling ILC Detector R&D Chris Damerell

  49. ISIS – Andrei Nomerotski Operating principles: • Charge collected under a photogate • Charge is transferred to 20-cell storage CCD in situ, 20 times during the 1 ms-long train • Conversion to voltage and readout in the 200 ms-long quiet period after the train(insensitive to beam-related RF pickup) • As in CCDs and pinned photodiode imaging pixels (aka 4 T pixels), the output gate decouples the charge collection from the charge sensing function, which can dramatically improve the noise performance • 1 MHz column-parallel readout is sufficient ILC Detector R&D Chris Damerell

  50. ISIS combines CCDs, in-pixel transistors and CMOS edge electronics in one device:non-standard process • Proof-of-principle device (ISIS1) designed and manufactured by e2V Technologies – works fine • ISIS2 (a prototype close to design goals) designed at RAL (Konstantin Stefanov and Pete Murray), due for delivery from Jazz Semiconductors any day now: • Modified 0.18 μm CMOS process with CCD-like buried channel and deep p+ implants. Single level (non-overlapping) poly for collection and transfer gates • Jazz have had success with mixed CMOS-CCD pixel structures, so we have some confidence … • Currently 80x10 mm storage pixel: goal is 80x5, leading to 20x20 imaging pixel as shown (slightly trapezoidal) • If too challenging, vertical integration can come to the rescue … Global Photogate and Transfer gate ROW 1: CCD clocks ROW 2: CCD clocks 80 mm On-chip switches On-chip logic ROW 3: CCD clocks ROW 1: RSEL Global RG, RD, OD RG RD OD RSEL Column transistor 5 μm ILC Detector R&D Chris Damerell

More Related